akshath raghav ravikiran

I’m a senior @ Purdue, pursuing a B.Sc in CompE.

My primary interest revolves arould building robust user-facing solutions at the intersection of explainable learning algorithms and interoperable systems. I enjoy research-driven environments aimed at taking concepts to tangible products.

Automatability, reproducibility and accessability remain the core of my work.

I’m concurrently working on multimodal document understanding, model-task alignment and integrated tools for AI agents. Post this, I intend to dive into accelerated computing and parallized ML architectures. Feel free to reach out to me at araviki[at]purdue[dot]edu.

background

• Smr’24: I’ve joined the Purdue SoCET group, in the Digital Design team. Over the summer, I added the Zicond Extension to the RISCV core for the AFTx07 tape-out. The goal is to reduce our branch mispredictions and simplify the conditional logic.
• Smr’24: I’m interned at VLSI System Design, a semiconductor ed-tech platform, where I implemented the TinySpeech family of speech recognition models for their vsdsquadronmini boad. I also wrote an ANSI-C based inference engine which runs out-of-box in 8bit precision, with 91%+ accuracies on the CH32V003F4U6 chip (2kb volatile sram, 16kb external mem).
• F’23 - S’24: I worked at the Duality Lab, where we re-engineered the MaskFormer segmentation model (funded by Google!) from the PyTorch-based artifact to TensorFlow for publishing to the TF Model Garden. You can find our paper here and code here. I also generated figures for the PeaTMOSS paper (accepted at MSR’24).
• S’24: I’m involved in MultiModal (LM) understanding projects at the e-lab. I’ve built eugenie & grammarflow. I am currently working on encoding reading-order patterns within documents (here’s my current plan of action), and building a routing-system for multi-modal RAG.
• S’24: I led a project at the CVES group @ Purdue ECE, where our goal was to define and evaluate reproducibility within AI/ML projects. I wrote the codebase for building our pipeline and statistically defined the importance of parameters.
• S’23 - Smr’23: I was employed at Ambee, where I deployed a worldwide fire forecasting system (F3) into their API and wrote automated scripts for their environment-data focused data lakes (still in use). You can find my LOR here. You can find the whitepaper here. You can get the data here.
• F’22 - S’23: I helped lead a project that was supervised by Prof. Yuan Wang (currently at Stanford) where we aimed to correlate lightning activity with wildfire spread. I wrote (big-)data-interfacing code for satellites across EUR/EUS/SAR, and was responsible for packing them to use within a ConvLSTM model from DeepCube’s short-term forecasting.

Find my reports here.

hobbies

In my free time, I enjoy photography, reading manga, watching @Geopold and speed typing (130+ wpm). At one point, I was religiously keeping up to date with 40 mangas concurrently.

On Sundays, I meal prep with my roommates; we alternate between cuisines regularly with Thai curries being our favorite. On Saturdays, I play Valorant with my friends till sunrise, or we make banana bread while watching old Telugu movies. During the week, I love cycling back to my dorm with friends and chugging apple juice (Mott’s is the best) during dinner.